Matrix in native Mojo with dynamic vector without numpy

[survuvi]

Like to python code below:

%%python

def create_matrix(rows, cols):
    # Create empy matrix
    matrix = []

    # Create matrix without numpy:
    for i in range(rows):
        row = []  # create a new row
        for j in range(cols):
            row.append(i*cols + j + 1)  # add one element into the row
        matrix.append(row)  # add the row into the matrix

    return matrix

def print_matrix(matrix):
    # Print matrix
    for i in range(len(matrix)):
        for j in range(len(matrix[0])):
            print(matrix[i][j], end=' ')
        print()  # start new line at the end of the row

# Print matrix 9 rows and 5 cols:
matrix = create_matrix(9, 5)
print_matrix(matrix)

'''
The output is:
[[ 1  2  3  4  5]
 [ 6  7  8  9 10]
 [11 12 13 14 15]
 [16 17 18 19 20]
 [21 22 23 24 25]
 [26 27 28 29 30]
 [31 32 33 34 35]
 [36 37 38 39 40]
 [41 42 43 44 45]]
'''

I'm trying to write a matrix in native mojo with dynamic vector without numpy:

from Vector import DynamicVector

fn create_matrix(rows : Int, cols : Int) raises -> object:
    var matrix = DynamicVector[Int]()
    for i in range(rows):
        var row = DynamicVector[Int]()
        for j in range(cols):
            row.push_back(i*cols + j + 1)
            matrix.push_back(row)
    return matrix

create_matrix(9,5)

_error: Expression [10]:12:29: invalid call to 'push_back': method argument #0 cannot be converted from 'DynamicVector[Int]' to 'Int'
matrix.push_back(row)
~~~~~~~~~~~~~~~~^~~~~

/.modular/Kernels/mojo/Stdlib/Vector.mojo:401:5: function declared here
fn push_back(inout self, value: type):
^

error: Expression [10]:13:12: cannot implicitly convert 'DynamicVector[Int]' value to 'object' in return value
return matrix_

[sa-]

You might be interested in checking out my Tensor implementation!
Warning, there's a lot of code and I will document it all much better when we have local dev
#251

[mfranzon]

@survuvi Why not use the struct Matrix as written in the documentation? https://docs.modular.com/mojo/notebooks/Matmul.html#adding-types-to-the-python-implementation

After create the structure like in the documentation you can use it in this way:

var C = Matrix(3, 3)
for i in range(C.rows):
    for j in range(C.cols):
        print(C[i,j])

[taalhaataahir0102]

Hi! Is there any way I can create an array or dynamic vector of objects?
e.g., This is my 1-D array:

struct Array1D[T: AnyRegType]:
    var data: Pointer[T]
    var dim0: Int

    fn __init__(inout self, dim0: Int):
        self.dim0 = dim0
        self.data = Pointer[T].alloc(dim0)
    
    fn __copyinit__(inout self, other: Array1D[T]):
        self.dim0 = other.dim0
        self.data = Pointer[T].alloc(self.dim0)
        for i in range(self.dim0):
            self.data.store(i, other.data.load(i))

    fn __getitem__(borrowed self, i: Int) -> T:
        return self.data.load(i)

    fn __setitem__(inout self, i: Int, value: T):
        self.data.store(i, value)

    fn __del__(owned self):
        self.data.free()

But it will only allow me to store data which could be fit in a register like int, float etc. I want to store objects in an array